DM-38062: Allow test to pass if --log-level is used for pytest

[timj]

There is still a confusing problem if pytest --log-cli-level is used.

Checklist

• ran Jenkins
• added a release note for user-visible changes to doc/changes

[codecov]

Codecov Report

Base: 85.51% // Head: 85.51% // No change to project coverage 👍

Coverage data is based on head (be6ae92) compared to base (f8dc3d2).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #790   +/-   ##
=======================================
  Coverage   85.51%   85.51%           
=======================================
  Files         266      266           
  Lines       35087    35087           
  Branches     7364     7364           
=======================================
  Hits        30004    30004           
  Misses       3768     3768           
  Partials     1315     1315           

Impacted Files	Coverage Δ
python/lsst/daf/butler/tests/cliLogTestBase.py	85.61% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[timj]

For the record, if I use --log-cli-level=DEBUG tests involving the Click runner mysteriously fail with:

>               stdout = outstreams[0].getvalue()
E               ValueError: I/O operation on closed file.

../../miniconda/envs/lsst-scipipe-5.0.0/lib/python3.10/site-packages/click/testing.py:434: ValueError

If I modify the click testing package to put a try/except around that line then all tests pass. I have no idea what is causing the problem (--log-cli-level=16 works but =15 fails).

Seems to be a known problem: pallets/click#824